The present invention relates to amplifier efficiency. More specifically, the invention provides techniques by which substantial improvements in amplifier efficiency may be realized without unacceptable increases in distortion.
Efficient power amplification is typically of fundamental importance in the operation of electrical and electronic systems. This is particularly true with regard to mobile communication systems. A great deal of the design effort for radio frequency (RF) communication devices is focused on improving the power efficiency of the RF amplifiers upon which such devices are based. Some designers have opted to base their devices on more efficient amplifier configurations such as, for example, class B, class E, or class F amplifiers. However, while such configurations are highly efficient, their nonlinear operation typically results in levels of out-of-band noise which are unacceptable because of the interference with adjacent communication channels.
On the other hand, amplifiers biased for class A or class AB operation do not, by definition, introduce such nonlinearities. Unfortunately, because the devices upon which such amplifiers are based are biased in their linear regions of operation, the efficiency with which such amplifiers operate is still undesirably low. This is exacerbated by the fact that mobile RF communication devices are typically required to operate well below their peak efficiency points. That is, most such devices have their greatest efficiency at higher output power levels. However, in a typical cellular device system, the closest cell transceiver receives the initial transmission from a cellular device and instructs the cellular device to back off on its output power level until a threshold level is reached below which the cell would not be able to detect transmissions from the device. That is, the cell transmits power control bits to the cellular device with which the device sets the output power level only to the level necessary to establish the link to the cell. This power management achieves the goal of maximizing the number of user which may be assigned to a given channel.
Some systems have employed bias control techniques to improve the power efficiency of RF amplifiers. In such systems, either or both of the bias voltage and bias current applied to the power output stage of an RF amplifier is manipulated with reference to the magnitude of the input RF signal to achieve a more efficient operation point than if the bias point was fixed. A discussion of bias control techniques as well as an example of a dual bias control technique is provided in a paper by Kyounghoon Yang et al. entitled High-Efficiency Class-A Power Amplifiers with Dual-Bias-Control Scheme, IEEE Transactions on Microwave Theory and Techniques, Vol. 47, No. 8, pp. 1426-1432, August 1999, the entirety of which is incorporated herein by reference for all purposes.
Unfortunately, there are significant limitations on how far bias techniques can go before nonlinearities are introduced which result in unacceptable distortion both inside and outside of the band of interest. That is, for example, when a cellular base station tells a particular device to transmit at a certain power level, this is typically achieved by setting the bias voltage for the RF amplifier at a certain level. This is due to the fact that the output power level is directly proportional to the square of the peak-to-peak excursion of the amplifier""s output which, in turn, is limited by the amplifier bias voltage. Therefore, the alternative for improving the power efficiency of the amplifier is to lower the bias current for the power transistor(s). However, the bias current for a typical device in an RF amplifier cannot go much below a certain level (which is device and technology dependent) before the distortion due to such nonlinearities reaches unacceptable levels.
In some amplifier designs, noise-shaping negative feedback techniques may be employed to reduce in-band distortion. Examples of such techniques are described in International Application No. PCT/US01/06780 for RF COMMUNICATION SYSTEM USING AN RF DIGITAL AMPLIFIER filed on Mar. 1, 2001, the entire disclosure of which is incorporated herein by reference for all purposes. Unfortunately, the most common configuration in RF amplifiers today is a single, common emitter/drain transistor with the collector/source coupled to a positive bias voltage through an inductor, a configuration which is not amenable to negative feedback techniques.
That is, the transistor in such amplifiers is biased well into its linear region of operation when conducting, but is typically off for much of the negative swing of the input RF signal being amplified (i.e., the inductor is conducting during this part of the cycle). Thus, the forward signal path of the amplifier becomes effectively an open circuit for roughly half of the signal cycle, introducing an effective delay which would undermine the stability of any negative feedback loop. That is, as is well known in feedback theory, a delay in a feedback loop results in an effective phase shift between the input signal and the feedback signal. When the delay is of sufficient magnitude and the phase shift approaches 180 degrees, the intended negative feedback becomes positive feedback and creates the potential for instability in the amplifier which typically takes the form of undamped oscillations.
In view of the foregoing, it is desirable to provide techniques by which the power efficiency of RF amplifiers may be improved while maintaining low distortion levels.
According to the present invention, an amplifier configuration is provided in which noise shaping feedback techniques are employed to reduce in-band distortion even where a feedback loop including the power switching device(s) is broken during some portion of the operational cycle. This is achieved by providing a feed forward path in parallel with a portion of the forward signal path of the amplifier which maintains loop integrity during all portions of amplifier operation. Even more generally, the feed forward path of the present invention may be employed to facilitate the use of negative feedback techniques in a wide variety of amplifier topologies for which such techniques were previously problematic to because of loop stability problems.
Thus, the present invention provides an amplifier comprising an amplification stage in a first feedback loop. The amplification stage is operable to open the first feedback loop during operation of the amplifier. The amplifier further comprises a feed forward path bypassing the amplification stage. The feed forward path is operable to provide a second feedback loop when the first feedback loop is open.
According to a more specific embodiment, an amplifier having a frequency band of interest associated therewith is provided. The amplifier comprises a frequency selective network, an amplification stage, and feedback circuitry in a first feedback loop. The frequency selective network is operable to provide noise shaping in the frequency band of interest using negative feedback via the feedback circuitry. The amplification stage is operable to open the first feedback loop during operation of the amplifier. The amplifier further comprises feed forward circuitry coupled to the first feedback loop and bypassing the amplification stage. The feed forward circuitry is operable to provide a second feedback loop when the first feedback loop is open thereby preserving amplifier loop stability.
According to an even more specific embodiment, a radio frequency (RF) amplifier having a frequency band of interest associated therewith is provided. The RF amplifier comprises a frequency selective network, an amplification stage, and feedback circuitry in a first feedback loop. The frequency selective network includes at least one resonator for providing noise shaping in the frequency band of interest using negative feedback via the feedback circuitry. The amplification stage comprises a switch and an inductor which are operable to alternately conduct load current. The switch is operable to open the first feedback loop while the inductor is conducting the load current. The RF amplifier further comprises feed forward circuitry coupled to the first feedback loop and bypassing the amplification stage. The feed forward circuitry is operable to provide a second feedback loop when the first feedback loop is open thereby preserving amplifier loop stability.
A further understanding of the nature and advantages of the present invention may be realized by reference to the remaining portions of the specification and the drawings.